A highly C70 selective shape-persistent rectangular prism constructed through one-step alkyne metathesis

Article abstract of DOI:10.1021/ja210418t

Dynamic covalent chemistry (DCC) provides an intriguing and highly efficient approach for building molecules that are usually thermodynamically favored. However, the DCC methods that are efficient enough to construct large, complex molecules, particularly those with three-dimensional (3-D) architectures, are still very limited. Here, for the first time, we have successfully utilized alkyne metathesis, a highly efficient DCC approach, to construct the novel 3-D rectangular prismatic molecular cage COP-5 in one step from a readily accessible porphyrin-based precursor. COP-5 consists of rigid, aromatic porphyrin and carbazole moieties as well as linear ethynylene linkers, rendering its shape-persistent nature. Interestingly, COP-5 serves as an excellent receptor for fullerenes. It forms 1:1 complexes with C₆₀ and C₇₀ with association constants of 1.4 × 10⁵ M^-1 (C₆₀) and 1.5 × 10⁸ M^-1 (C₇₀) in toluene. This represents one of the highest binding affinities reported so far for purely organic fullerene receptors. COP-5 shows an unprecedented high selectivity in binding C₇₀ over C₆₀ (K_C70/K_C60 > 1000). Moreover, the binding between the cage and fullerene is fully reversible under the acid-base stimuli, thus allowing successful separation of C₇₀ from a C₆₀-enriched fullerene mixture (C₆₀/C₇₀, 10/1 mol/mol) through the "selective complexation-decomplexation" strategy.

Full text of DOI:10.1021/ja210418t

ARTICLE
pubs.acs.org/JACS
A Highly C70 Selective Shape-Persistent Rectangular Prism Constructed
through One-Step Alkyne Metathesis
Chenxi Zhang,† Qi Wang,† Hai Long,‡ and Wei Zhang*,†
†Department of Chemistry and Biochemistry, University of Colorado, Boulder, Colorado 80309, United States
‡National Renewable Energy Laboratory (NREL), Golden, Colorado 80401, United States
S Supporting Information
b
ABSTRACT:
Dynamic covalent chemistry (DCC) provides an intriguing and highly efficient approach for building molecules that are usually
thermodynamically favored. However, the DCC methods that are efficient enough to construct large, complex molecules,
particularly those with three-dimensional (3-D) architectures, are still very limited. Here, for the first time, we have successfully
utilized alkyne metathesis, a highly efficient DCC approach, to construct the novel 3-D rectangular prismatic molecular cage COP-5
in one step from a readily accessible porphyrin-based precursor. COP-5 consists of rigid, aromatic porphyrin and carbazole moieties
as well as linear ethynylene linkers, rendering its shape-persistent nature. Interestingly, COP-5 serves as an excellent receptor for
fullerenes. It forms 1:1 complexes with C60 and C70 with association constants of 1.4 Â 105 MÀ1 (C60) and 1.5 Â 108 MÀ1 (C70) in
toluene. This represents one of the highest binding affinities reported so far for purely organic fullerene receptors. COP-5 shows an
unprecedented high selectivity in binding C70 over C60 (KC70/KC60 > 1000). Moreover, the binding between the cage and fullerene
is fully reversible under the acidÀbase stimuli, thus allowing successful separation of C70 from a C60-enriched fullerene mixture
(C60/C70, 10/1 mol/mol) through the “selective complexationÀdecomplexation” strategy.
hree-dimensional (3-D) molecular cages, particularly shape-
persistent, covalent organic polyhedrons (COPs) with well-
imine condensation/metathesis is almost the only reversible
DCC reaction that has been used in the construction of 3-D
molecular architectures.8,9,24,25 However, the potential draw-
backs of imine groups are their sensitivity to acidic conditions
and water. Further hydride reduction of imines provides more
robust, but also flexible, amino groups, resulting in the loss of
certain shape persistency of target structures.
T
defined pore dimensions, have attracted considerable attention
due to their numerous applications in hostÀguest chemistry,1,2
chemical sensing,3,4 catalysis,5,6 gas adsorption,7À10 etc. Current
synthesis of rigid molecular cages is dominated by supramolec-
ular chemistry, including metal coordination6,11À13 and hydro-
gen bonding,14 which usually provides the target species with
high efficiency through the self-assembly process. The growing
importance of supramolecular cages in the bottom-up design of
novel functional materials has prompted the development of
COPs, which usually possess high chemical and thermal stability,
as promising candidates for those applications requiring harsh
operating conditions (e.g., in the presence of acid or base). While
supramolecular cages have been extensively studied, purely
organic covalent molecular cages are relatively rare and have
only recently received increasing attention. Conventionally,
COPs are constructed via irreversible chemical transformations,
which usually require enormous synthetic and purification efforts
with very low overall yields.15,16 In great contrast, recent ad-
vances in dynamic covalent chemistry (DCC)17,18 offer conve-
nient pathways to high-yielding synthesis of COPs.19À24 To date,
Rigid, shape-persistent 3-D molecular cages26 are attractive as
hosts for selectively binding guest molecules of a certain size and
shape. However, the construction of molecular cages with well-
defined and large intrinsic cavities, which are not easily altered
by the environmental stimuli, still remains a grand challenge.
To preserve such shape persistency, the molecules should
consist of rigid, noncollapsible scaffolds. In the ocean of con-
formationally rigid building blocks exists a particularly interest-
ing group of conjugated systems, porphyrins, which have attracted
enormous research interest due to their intriguing applications in
hostÀguestchemistry,27 artificialphotosynthesis,28 andcatalysis.29
Another important group of building blocks with inherent rigidity
Received: October 7, 2011
Published: November 15, 2011
r
2011 American Chemical Society
20995
dx.doi.org/10.1021/ja210418t J. Am. Chem. Soc. 2011, 133, 20995–21001
|

Journal of the American Chemical Society
ARTICLE
has been successfully utilized in the efficient synthesis of shape-
persistent 2-D aryleneethynylene macrocycles,37 we envisioned
this dynamic covalent transformation with “self-correction”
nature would be an ideal approach to construct the ethynyl-
ene-linked rigid molecular cage. The next challenge for the
dynamic covalent construction of a molecular cage was then
the identification of a rigid 90° corner piece that would yield a
rectangular-prismatic structure with minimum angle strain. Care-
ful examination of the geometrical parameters suggested that the
C4-symmetric carbazole-substituted porphyrin 3, in which the
ethynyl groups are oriented approximately perpendicular to the
porphyrin ring, could be a suitable monomer to yield COP-5
(Scheme 1). Porphyrins are incorporated into the cage structure
because of their interesting properties in fullerene binding, which
would provide highly efficient donorÀacceptor pairs for the
photoinduced electron transfer (PET) process.39,40
Figure 1. Schematic presentation of constructing a rectangular prism
from a tetra-armed precursor through alkyne metathesis.
are arylene ethynylenes, and recent advances in shape-persistent
2-D macrocycles have taken advantage of their angular, rigid
geometries and easy synthetic access.30À33 In this context, alkyne
metathesis,34À36 a DCC approach, has emerged as a powerful
synthetic tool for the efficient and versatile construction of shape-
persistent molecular architectures and has been elegantly applied
to the one-step modular synthesis of a variety of 2-D macrocyclic
compounds from simple arylene diacetylene monomers.37 Such a
synthetic approach is remarkable in that it can transform various
undesired oligomeric and macrocyclic kinetic intermediates into
the thermodynamically most stable product. Shape-persistent
3-D molecular architectures based on rigid arylene ethynylene
backbones are synthetically more challenging than their 2-D ana-
logues, and so far they have mainly been obtained through
multistep labor-intensive statistical or template-supported oxida-
tive Glaser coupling.38 Here we present, for the first time, one-
step synthesis of a shape-persistent organic rectangular prismatic
cage consisting of only aromatic moieties in its backbone
structure, via Mo(VI) carbyne catalyzed alkyne metathesis
from tetrasubstituted porphyrin monomers, in decent yield
(Figure 1). The unique combination of the conjugated system
and the rigidity of this cage molecule renders its rapid and
selective binding of C70 over C60; the cage exhibits a binding
interaction with C70 compared to C60 that is stronger by
3 orders of magnitude. Moreover, the clean release of ful-
lerenes (guest) and regeneration of the cage COP-5 (host)
was realized by simply tuning the pH of the media. Such a
“selective complexationÀdecomplexation” strategy has been
successfully applied to the isolation of C70 from a C60-enriched
fullerene mixture.
The monomer 3 was prepared from 3-iodo-6-formyl-9-
hexadecylcarbazole (1) through the Lindsey method to form the
5,10,15,20-tetrakis(carbazolyl)porphyrin 2, followed by Sonoga-
shira coupling to attach the benzoylbiphenyl acetylene group
(Scheme 1). Benzoylbiphenyl was utilized as the end group so
that insoluble byproduct diarylacetylenes would be formed along
the reaction, thus driving the reversible alkyne metathesis to
completion.37 Our newly developed multidentate Mo(VI) car-
byne was utilized as the metathesis catalyst, which has shown
superior catalytic activity, broader substrate scope, and longer
lifetime in comparison to its monodentate analogue.41 Given the
suitable geometrical shape, rigidity of the building block, and
high activity of alkyne metathesis catalyst, the formation of the
thermodynamically favored target molecular cage is expected.
The reaction was performed at 75 °C under microwave irradia-
tion in CCl4. After 32 h, the predominant formation of cage
COP-5 was observed (56% isolated yield). The dynamic nature
of alkyne metathesis offers the self-correction pathway toward
the cage formation. The reaction progress monitored by gel
permeation chromatography (GPC) showed the gradual con-
version of initial high-molecular-weight oligomeric intermediates
into the target cage COP-5 (Figure S2, Supporting Information).
Such a cage is enthalpy favored due to its minimal angle strain
and also entropy favored due to its consistent minimal number of
building blocks (in comparison to larger oligomeric products).
This cage synthesis represents the first time alkyne metathesis
has been used to construct a purely organic, shape-persistent
molecular rectangular prism with high efficiency.
The cage COP-5 was fully characterized by 1H and 13C NMR
spectroscopy, UVÀvis spectroscopy, and GPC, as well as
’ RESULTS AND DISCUSSION
1
Previously our group has demonstrated that various trigonal-
prismatic cage molecules can be constructed through the “panel-
pillar-assembly” design strategy via a one-step imine condensa-
tion/metathesis reaction followed by further reduction of imines
to more chemically robust amino groups.9 However, these
amine-linked molecular cages are semishape-persistent and
adopt a distorted-prismatic architecture due to the rotational
freedom of the amine bond connections. Although COPs con-
taining methylene, ether, amino, or similar saturated moieties
could be shape-persistent to some extent, they are generally
considered less conformationally rigid compared to those
molecules constructed solely with aromatic fragments in their
skeletons. Among functional groups capable of undergoing
chemical transformations while retaining their conformational
rigidity, the linear ethynylene group seems to be an appropriate
robust rigid linker moiety. Given the fact that alkyne metathesis
MALDI-TOF mass spectrometry. The H NMR spectrum of
COP-5 in CDCl3 shows only one set of singlets corresponding
to the porphyrin protons at 8.73 ppm, indicating the high
symmetry of the cage structure. The MALDI-TOF mass
spectrum shows the desired molecular ion peaks at m/z
3825.80 ([M + H]+ calcd for C272H332N16 3825.66). Additional
evidence supporting the rectangular-prismatic structure of
COP-5 was obtained through a diffusion ordered spectroscopy
(DOSY) experiment. All the peaks assigned to COP-5 have the
same diffusion coefficients (D values) within the experimental
errors ((4.1 ( 0.4) Â 10À10 m2 sÀ1). From the StokesÀ
Einstein equation, the effective hydrodynamic radius R
(9.2 ( 1.0 Å) is calculated, which is in good agreement with
the calculated radius of gyration 9.7 Å based on the energy-
minimized structure in Figure 2. The cage is thermally stable
and also exhibits a very high chemical stability even with
20996
dx.doi.org/10.1021/ja210418t |J. Am. Chem. Soc. 2011, 133, 20995–21001

Journal of the American Chemical Society
ARTICLE
Scheme 1. Synthesis of Rectangular Prism COP-5
exposure to water and acids (e.g., trifluoroacetic acid, TFA) for
weeks, thus showing a great advantage over those supramolec-
ular cages as well as imine-linked COPs.
by the substantial intensity decrease (∼55%, with 1 equiv of C70
added) at 428 nm and the red shift (9 nm) of the porphyrin Soret
band in comparison to that of COP-5 itself. A similar trend in the
UVÀvis titration curve was also observed when the solution of
COP-5 in toluene was titrated with C60. According to the Job
plot, both C60 and C70 formed a 1:1 hostÀguest complex with
COP-5 (Figure S4, Supporting Information). Consistent with
this observation, when 0.25 equiv of C70 was added to the cage
solution in toluene, C70@COP-5 was formed instantaneously,
and the 1H NMR spectra (Figure 3c) showed two sets of signals
corresponding to the free COP-5 and C70@COP-5 complex
with an integration ratio of 3:1, further supporting the 1:1
binding mode between COP-5 and C70. MALDI-TOF spectra
(Figure S5, Supporting Information) of cageÀfullerene com-
plexes clearly showed peaks with m/z 4544.20 and 4664.74
corresponding to 1:1 hostÀguest complexes C60@COP-5 and
C70@COP-5, without any other complexes observed.
Additional evidence in support of the fullerene encapsulation
inside the cage comes from the analysis of 1H NMR spectra of the
cageÀfullerene complexes. The chemical shifts of the protons at
the 4,5-positions on the carbazole corner pieces, which are
pointing to the inside cavity of the cage, moved significantly
downfield in both C60@COP-5 and C70@COP-5, while the
other protons of the carbazole are not much affected (Figure S6,
Supporting Information); such an observation indicates the
fullerenes are located inside the cage. On the other hand, the
internal NÀH protons of the porphyrins at À1.97 ppm shifted
One of our major motivations for constructing such rigid,
porphyrin-containing molecular cages is to develop a novel class
of organic fullerene receptors, which could bind strongly and also
selectively with a certain type of fullerene. Although covalent
macrocycles27,42 or supramolecular cages43 containing porphyrin
moieties have been reported to exhibit strong binding interac-
tions with fullerenes, covalently bonded organic cages have rarely
been reported as fullerene receptors. PorphyrinÀfullerene bind-
ing is mainly driven by an electronic effect: i.e., a favored
donorÀacceptor interaction. With two porphyrin-free base
moieties as the top and bottom panels, the cage COP-5 is
expected to strongly bind fullerenes. The computational model-
ing study (Figure 2) reveals that COP-5 has a cavity with a height
(defined as the distance between the top and bottom porphyrin
panels) of 11.9 Å and a diameter of 18.3 Å (defined as the
distance between the two ethynylene groups in the diagonal
edges), in which a fullerene can be nicely accommodated.
The cage COP-5 indeed showed a strong binding interaction
with fullerenes. The binding of COP-5 with C60 and C70 was
characterized by UVÀvis titration experiments in toluene
(Figure 3a,b). With a gradual addition of C70 to the cage solution
(in toluene), the intensity of the absorption peak of COP-5 at
428 nm decreases while a new signal at 437 nm arises. The
cageÀC70 complex (C70@COP-5) formation is clearly signaled
20997
dx.doi.org/10.1021/ja210418t |J. Am. Chem. Soc. 2011, 133, 20995–21001

Journal of the American Chemical Society
ARTICLE
Figure 2. Energy-minimized structures of COP-5 (a, top view; b, side
view), C70@COP-5 (c), and C60@COP-5 (d). Methyl groups were
used in the calculation instead of hexadecyl chains for simplicity. The
height of COP-5 was defined as the distance between the top and
bottom porphyrin panels, and the diameter of the inside cavity of COP-5
was defined as the distance between the two ethynylene groups in the
diagonal edges.
significantly upfield by the influence of the fullerene ring current
in C70@COP-5 (Figure S6), which indicates a strong πÀπ
interaction between C70 and the porphyrin panels of COP-5.
The simplicity of the NMR signals of the C70@COP-5 and
C60@COP-5 suggests a highly symmetrical structure of the
complexes, further supporting the notion of fullerene binding
inside the cage.
On the basis of the 1:1 binding mode and fitting of the UVÀvis
adsorption changes at 428 nm under different fullerene concen-
trations, the association constants of C60 and C70 with COP-5
were estimated to be 1.4 Â 105 MÀ1 (C60) and 1.5 Â 108 MÀ1
(C70) in toluene, which are comparable to those best-performing
fullerene receptors reported thus far. It has been a great challenge
to obtain host molecules that bind fullerenes with association
constants higher than 105 MÀ1. To the best of our knowledge,
the highest fullerene binding affinity was observed in a Ir(III)
metalloporphyrin derivative reported by the Aida group, which
showed Kassoc = 1.3 Â 108 MÀ1 in 1,2-dichlorobenzene and
Kassoc > 109 MÀ1 in benzene for C60.44 The Anderson group also
reported a cyclic porphyrin trimer with an impressive high
Figure 3. COP-5Àfullerene (C70 and C60) binding studies. (a) UVÀvis
absorption spectra of COP-5 (2.0 μM) in toluene in the presence of
various amounts of C70 (0 f 3 equiv) at 23 °C, while the concentration
of COP-5 was maintained constant. Inset: plot of ΔA428
vs
nm
equivalents of C70 added. (b) UVÀvis absorption spectra of COP-5
(2.0 μM) in toluene in the presence of various amounts of C60 (0 f 50
equiv) at 23 °C, while maintaining the concentration of COP-5 constant.
Inset: plot of ΔA428 nm vs equivalents of C60 added. The association
constants modeled with a 1:1 equilibrium are KC70@COP‑5 = 1.5 Â 108 MÀ1
(ΔG=À11.2 kcal/mol) for C70@COP-5 and KC60@COP‑5 = 1.4Â 105 MÀ1
(ΔG = À7.0 kcal/mol) for C60@COP-5. (c) 1H NMR spectra of
C70@COP-5 (blue), COP-5 (green), and a mixture of COP-5 with
0.25 equiv of C70 (red) in toluene-d8.
affinity for fullerenes with Kassoc = 2 Â 106 for C60 and Kassoc
=
2 Â 108 for C70 in toluene.42 However, it should be noted that the
previously reported best performers are all metalloporphyrins
and release of the fullerene guest and regeneration of the valuable
host would be difficult, thus making the fullerene separation/
purification based on these host molecules impractical.
of COP-5 with C70 in a C60-enriched fullerene mixture was
observed. When COP-5 was mixed with a solution of C60
(91 mol %) and C70 (9 mol %) in toluene, the COP-5 selectively
bound with C70 to form C70@COP-5. The 1H NMR spectrum
clearly shows the major set of proton signals corresponding to
C70@COP-5 (Figure S6).
Computational calculations on the energy-minimized struc-
tures of COP-5 and COP-5Àfullerene complexes provide
further insight into the preferential binding of C70 versus C60.
The Amber 11.0 molecular dynamics program package45 was
used to optimize the structure of the fullerene, the cage, and the
cageÀfullerene binding complexes. The general Amber force
field (GAFF field)46 was used with the charge parameters
computed by the AM1-BCC method.47 The computational
It is noteworthy that the cage COP-5, containing nonmeta-
lated porphyrin moieties, shows a high binding affinity with C70,
which is 3 orders of magnitude higher than that with C60. Such a
remarkable selectivity in the binding of C70 over C60 is unprec-
edented and is significantly higher than those of previously
reported fullerene receptors, which are usually lower than 10,
with highest being around 100, reported by the Anderson
group.42 In order to further explore the potential of COP-5
in fullerene separation, we used a mixture of two fullerene guests in
a binding competition test. As expected, selective complexation
20998
dx.doi.org/10.1021/ja210418t |J. Am. Chem. Soc. 2011, 133, 20995–21001

Journal of the American Chemical Society
ARTICLE
modeling study (Figure 2) reveals that the heights of COP-5 are
slightly increased to 12.1À12.2 Å in C60@COP-5 and
C70@COP-5 from the initial 11.9 Å in the unoccupied COP-5,
while the diameters of the cavities are decreased slightly to
17.6À17.8 Å from 18.3 Å. Owing to the high degree of shape
persistency of the cage COP-5, the size and geometry of the cage
do not change much upon fullerene encapsulation. In both
C60@COP-5 and C70@COP-5 complexes, the shortest atom-
to-atom distances between porphyrin panels and fullerenes are
similar and are close to 3.2 Å, which leads to an appreciable πÀπ
interaction in both complexes. It is known that fullerenes also
have CHÀπ interactions with host molecules, and the favored
distance for CHÀπ interactions is around 2.9 Å.48,49 This type of
CHÀπ interactions also exists in fullereneÀCOP-5 complexes.
The average distances between fullerenes and those carbazole
CH protons, which are pointing toward the inside cavity of the
cage, in C60@COP-5 and C70@COP-5 are around 3.5 and 3.1 Å,
respectively. Since the distance for CHÀπ interactions in
C70@COP-5 (3.1 Å) is closer to the known favored distance
2.9 Å, the CHÀπ interactions are presumably stronger in
C70@COP-5 compared to those in C60@COP-5. Considering
the rigid, noncollapsible nature of the cage COP-5, it would be
difficult for the cage to adjust the distance between the top and
bottom porphyrin panels to achieve stronger binding interac-
tions with fullerene guests. The nonperfect match between the
sizes of host and guest would result in a less favored binding
interaction of COP-5 with C60. Given the conformational
rigidity, the slight difference in the distance between the host
and guest could lead to a large difference in their stabilization
energy (4.2 kcal/mol, calculated from the binding constants of
C70@COP-5 and C60@COP-5) gained through the hostÀguest
binding interaction and could be the origin of the remarkable
binding selectivity.
Figure 4. pH-Driven reversible COP-5Àfullerene binding. COP-5 con-
centration was 1.0 Â 10À6 M in toluene. (a) UVÀvis spectra of the free
COP-5 (cyan), ofC70@COP-5 (magenta), after theaddition of100equiv
of TFA to COP-5 (black), after the addition of 100 equiv of TFA to
C70@COP-5 (red), and after the addition of 100 equiv of TFA, followed
by the subsequent addition of 100 equiv of TEA to C70@COP-5 (blue).
Inset: plot of absorption at 437 nm vs repetitive associationÀdissociation
cycles. (b) UVÀvis spectra of the free COP-5 (cyan), of C60@COP-5
(magenta), after the addition of 100 equiv of TFA to COP-5 (black), after
the addition of 100 equiv of TFA to C60@COP-5 (red), and after the
addition of 100 equiv of TFA followed by the subsequent addition of 100
equiv of TEA to C60@COP-5 (blue). Inset: plot of absorption at 436 nm
vs repetitive associationÀdissociation cycles.
Intrigued by the impressive differentiating capability of cage
COP-5 in binding C60 and C70, and to achieve a clean and
efficient separation of C70 from fullerene mixtures, we then
directed our efforts to the investigation of controlled release
of fullerene guests and regeneration of host cage molecules.
Since porphyrinÀfullerene interactions are based on the attrac-
tion between electron-rich (porphyrin) and electron-poor
(fullerene) moieties, their interactions could be tuned by
changing the electronic properties of either one of them. Unlike
the case for metalloporphyrins, the electron density of the
porphyrin free base can be easily reduced by simple proton-
ation, and thus the porphyrinÀfullerene interactions could be
weakened. We envisioned dissociation and release of the guest
molecules, and regeneration of the COP-5 can be realized by
simply tuning the pH of the media. Trifluoroacetic acid (TFA)
and triethylamine (TEA) were used as the acid and base stimuli.
Upon addition of excess TFA (100 equiv) to a solution of
C70@COP-5 (or C60@COP-5) in toluene, protonation of the
porphyrin ring occurred, and consequently, complete release of
the fullerene molecules was observed, as evidenced by the
disappearance of the 1H NMR signals corresponding to
C70@COP-5 (or C60@COP-5) and the appearance of a new
set of signals corresponding to the protonated COP-5 with an
empty cavity. However, the subsequent addition of 100 equiv of
triethylamine (TEA) to the above mixture neutralized the
porphyrin ring and restored the binding interaction between
reversible association/dissociation triggered by acid/base stim-
uli was also confirmed by monitoring the process with
UVÀvis absorption spectra (Figure 4). In brief, we observed
almost the identical UVÀvis absorption spectra when we
protonated COP-5 and C70@COP-5 (or C60@COP-5) with
100 equiv of TFA, which indicates the complete dissociation
of the cageÀfullerene complex upon protonation of the por-
phyrin-free base center. The regeneration of C70@COP-5
(or C60@COP-5) complex by the addition of 100 equiv of
TEA to the mixture of protonated COP-5 and free C70 was
evidenced by a UVÀvis absorption almost identical with that of
the pure C70@COP-5 (or C60@COP-5). The associationÀ
dissociation process can be repeated in several cycles without
leading to any noticeable change in the absorbance by the
alternating addition of TFA and TEA (Figure 4, inset), thus
showing that the cageÀC70 binding is a highly efficient, robust,
and fully reversible process.
1
COP-5 and fullerenes. As a result, the H NMR spectrum of
Fullerene purification has attracted great attention since the
1990s, and people are still actively pursuing a highly efficient and
easy process for fullerene separations.42,50À52 One of the most
C70@COP-5 (or C60@COP-5) was regenerated, indicating
the reversibility of the association/dissociation process. Such a
20999
dx.doi.org/10.1021/ja210418t |J. Am. Chem. Soc. 2011, 133, 20995–21001

Journal of the American Chemical Society
ARTICLE
’ CONCLUSIONS
In summary, we have successfully synthesized the shape-
persistent porphyrin-based molecular cage COP-5 via efficient
alkyne metathesis, which represents the first time such a rever-
sible dynamic covalent approach has been used to construct 3-D
molecular cages. Such a DCC approach complements the exist-
ing dynamic assembly strategies for the construction of large
complex, molecular architectures and would be particularly
beneficial for the synthesis and study of noncollapsible, shape-
persistent covalent cage molecules. The obtained cage COP-5
shows an unprecedented high selectivity (>1000/1) in binding of
C70 over C60. The pH-driven reversible association and dissocia-
tion of fullereneÀcage complexes enables an easy and efficient
separation of C70 from a fullerene mixture. Currently the binding
behavior of such porphyrin-based molecular cages with higher
fullerenes and further assembly of these cageÀfullerene com-
plexes into highly ordered 3-D architectures are being investi-
gated in our laboratory and will be reported in due course.
Figure 5. Schematic presentation of the C70 isolation process. Step I: to
a solution of C60 and C70 mixture (C60/C70 = 10/1, molar ratio) in CS2
was added a small amount of COP-5 (equal to or less than the
stoichiometric amount of C70), resulting in the favored formation of
C70@COP-5. After the unbound free fullerenes were separated by
precipitation in CHCl3 (precipitates shown in (b)), cageÀfullerene
complexes (mostly C70@COP-5) were collected in the solution phase.
Step II: upon the addition of 100 equiv of TFA to the solution collected
in step II, fullerene guest molecules (mostly C70) were released as black
precipitates and removed to complete one cycle of the isolation process
(shown in (c)). Step III: regeneration of COP-5 was accomplished by
the addition of 100 equiv of TEA to the above solution.
’ ASSOCIATED CONTENT
S
Supporting Information. Text and figures giving experi-
b
mental details, UVÀvis spectroscopic data, computation meth-
ods, and NMR spectra. This material is available free of charge via
the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
common methods aiming at the separation of fullerenes of
specific size is based on selective complexation of a particular
fullerene from an organic solution of fullerene mixtures.27,53
However, the difficult release of fullerenes and regeneration of
hosts have greatly impeded the practical applications of this
method.52 Given the over 1000 times stronger binding interac-
tions of COP-5 with C70 over C60, and the reversible nature of
this hostÀguest binding triggered by pH, we envisioned that the
separation of C70 from C60 can be achieved with a simple
“selective complexationÀdecomplexation” process (Figure 5).
As a proof of concept, we used a C60-enriched C60/C70 mixture
in the separation study. Carbon disulfide was chosen as the
solvent for the encapsulation step, since both C60 and C70 have
good solubility in CS2. A mixture of COP-5 and C60/C70 in CS2
was sonicated for 30 s, and the solvent was evaporated. The
residue was then dispersed in chloroform. Since fullerenes have
very limited solubility in CHCl3, free fullerenes remained as
precipitates. The solution phase, which was composed of mostly
C70@COP-5, was separated from the insoluble fullerene mix-
tures by centrifugation. Further acidifying the C70@COP-5
complex with excess TFA followed by sonication (5 min)
released C70 as a black precipitate, which allowed easy removal
by centrifugation. The cage COP-5 was then regenerated by
neutralization with TEA and was recycled for the next round of
fullerene separation. The UVÀvis absorption showed that the
C70 abundance of the fullerene mixture increased significantly
(∼9-fold increase) from the initial 9 mol % to 79 mol % after only
one cycle of separation (Figure S1, Supporting Information).
This result clearly demonstrates the simplicity and high efficiency
of such a fullerene separation approach by using shape-persistent
molecular cages as selective receptors. The “selective complexationÀ
decomplexation” strategy presented here would greatly facilitate
the purification of these intriguing graphitic materials and pro-
mote their wide application in organic photovoltaics, polymer
electronics, and biopharmaceuticals.
Corresponding Author
wei.zhang@colorado.edu.
’ ACKNOWLEDGMENT
This work was supported by the National Science Foundation
(Nos. DMR-1055705 and CBET-1033255) and the Defense
Advanced Research Projects Agency (DARPA). We thank
Dr. Kuthanapillil Jyothish for helpful discussions, Dr. Yinghua
(Alice) Jin for help with manuscript writing, and Dr. Richard
Shoemaker for NMR assistance. This research used the capabilities
of the National Renewable Energy Laboratory Computational
Science Center, which is supported by the Office of Energy
Efficiency and Renewable Energy of the U.S. Department of Energy
under Contract No. DE-AC36-08GO28308.
’ REFERENCES
(1) Rebek, J., Jr. Angew. Chem., Int. Ed. 2005, 44, 2068–2078.
(2) Meissner, R. S.; Rebek, J., Jr; DeMendoza, J. Science 1995, 270,
1485–1488.
(3) Bisson, A. P.; Lynch, V. M.; Monahan, M. K. C.; Anslyn, E. V.
Angew. Chem., Int. Ed. 1997, 36, 2340–2342.
(4) Ferrand, Y.; Crump, M. P.; Davis, A. P. Science 2007, 318,
619–22.
(5) Thomas, J. M.; Raja, R.; Sankar, G.; Bell, R. G. Acc. Chem. Res.
2001, 34, 191–200.
(6) Inokuma, Y.; Kawano, M.; Fujita, M. Nat. Chem. 2011, 3, 349–358.
(7) Wang, B.; Cote, A. P.; Furukawa, H.; O’Keeffe, M.; Yaghi, O. M.
Nature 2008, 453, 207–211.
(8) Tozawa, T.; et al. Nat. Mater. 2009, 8, 973–978.
(9) Jin, Y. H.; Voss, B. A.; Noble, R. D.; Zhang, W. Angew. Chem., Int.
Ed. 2010, 49, 6348–6351.
(10) Atwood,J.L.;Barbour,L. J.;Jerga,A.Science 2002, 296,2367–2369.
(11) Olenyuk, B.; Whiteford, J. A.; Fechtenkotter, A.; Stang, P. J.
Nature 1999, 398, 796–799.
21000
dx.doi.org/10.1021/ja210418t |J. Am. Chem. Soc. 2011, 133, 20995–21001

Journal of the American Chemical Society
ARTICLE
(12) Seidel, S. R.; Stang, P. J. Acc. Chem. Res. 2002, 35, 972–983.
(13) Fiedler, D.; Leung, D. H.; Raymond, K. N.; Bergman, R. G. Acc.
Chem. Res. 2005, 38, 349–358.
(14) Liu, Y. Z.; Hu, C. H.; Comotti, A.; Ward, M. D. Science 2011,
333, 436–440.
(15) Ashton, P. R.; Isaacs, N. S.; Kohnke, F. H.; Dalcontres, G. S.;
Stoddart, J. F. Angew. Chem., Int. Ed. 1989, 28, 1261–1263.
(16) Ashton, P. R.; Girreser, U.; Giuffrida, D.; Kohnke, F. H.;
Mathias, J. P.; Raymo, F. M.; Slawin, A. M. Z.; Stoddart, J. F.; Williams,
D. J. J. Am. Chem. Soc. 1993, 115, 5422–5429.
(49) Umezawa, Y.; Tsuboyama, S.; Honda, K.; Uzawa, J.; Nishio, M.
Bull. Chem. Soc. Jpn. 1998, 71, 1207–1213.
(50) Huerta, E.; Metselaar, G. A.; Fragoso, A.; Santos, E.; Bo, C.;
de Mendoza, J. Angew. Chem., Int. Ed. 2007, 46, 202–205.
(51) Atwood, J. L.; Koutsantonis, G. A.; Raston, C. L. Nature 1994,
368, 229–231.
(52) Komatsu, N. J. Inclusion Phenom. Mol. Recognit. Chem. 2008, 61,
195–216.
(53) Martin, N.; Perez, E. M. Chem. Soc. Rev. 2008, 37, 1512–1519.
(17) Rowan, S. J.; Cantrill, S. J.; Cousins, G. R. L.; Sanders, J. K. M.;
Stoddart, J. F. Angew. Chem., Int. Ed. 2002, 41, 898–952.
(18) Lehn, J. M. Science 2002, 295, 2400–2403.
(19) Lehn, J. M.; Eliseev, A. V. Science 2001, 291, 2331–2332.
(20) Chichak, K. S.; Cantrill, S. J.; Pease, A. R.; Chiu, S. H.;
Cave, G. W. V.; Atwood, J. L.; Stoddart, J. F. Science 2004, 304, 1308–
1312.
(21) Sanders, J. K. M.; Corbett, P. T.; Leclaire, J.; Vial, L.; West,
K. R.; Wietor, J. L.; Otto, S. Chem. Rev. 2006, 106, 3652–3711.
(22) Hartley, C. S.; Moore, J. S. J. Am. Chem. Soc. 2007, 129,
11682–11683.
(23) Granzhan, A.; Schouwey, C.; Riis-Johannessen, T.; Scopelliti, R.;
Severin, K. J. Am. Chem. Soc. 2011, 133, 7106–7115.
(24) Liu, X. J.; Liu, Y.; Li, G.; Warmuth, R. Angew. Chem., Int. Ed.
2006, 45, 901–904.
(25) Meyer, C. D.; Joiner, C. S.; Stoddart, J. F. Chem. Soc. Rev. 2007,
36, 1705–1723.
(26) Mastalerz, M. Angew. Chem., Int. Ed. 2010, 49, 5042–5053.
(27) Tashiro, K.; Aida, T. Chem. Soc. Rev. 2007, 36, 189–197.
(28) Wasielewski, M. R. Chem. Rev. 1992, 92, 435–461.
(29) Meunier, B. Chem. Rev. 1992, 92, 1411–1456.
(30) Eickmeier, C.; Junga, H.; Matzger, A. J.; Scherhag, F.; Shim, M.;
Vollhardt, K. P. C. Angew. Chem., Int. Ed. 1997, 36, 2103–2108.
(31) Haley, M. M.; Pak, J. J.; Brand, S. C. Carbon Rich Compounds II;
Springer: Berlin, 1999; Topics in Current Chemistry 201, pp 81À130
(32) Adams, R. D.; Ge, P. H.; Fu, W.; Herrmann, W. A.; Herdtweck,
E.; Campana, C.; Bunz, U. H. F. Angew. Chem., Int. Ed. 2000, 39,
3607–3610.
(33) Zhang, W.; Moore, J. S. Angew. Chem., Int. Ed. 2006, 45,
4416–4439.
(34) Schrock, R. R. Chem. Rev. 2002, 102, 145–179.
(35) Furstner, A.; Davies, P. W. Chem. Commun. 2005, 2307–2320.
(36) Jyothish, K.; Zhang, W. Angew. Chem., Int. Ed. 2011, in press.
(37) Zhang, W.; Moore, J. S. J. Am. Chem. Soc. 2004, 126,
12796–12796.
(38) Bunz, U. H. F.; Rubin, Y.; Tobe, Y. Chem. Soc. Rev. 1999,
28, 107–119.
(39) Boyd, P. D. W.; Hodgson, M. C.; Rickard, C. E. F.; Oliver, A. G.;
Chaker, L.; Brothers, P. J.; Bolskar, R. D.; Tham, F. S.; Reed, C. A. J. Am.
Chem. Soc. 1999, 121, 10487–10495.
(40) Baskaran, D.; Mays, J. W.; Zhang, X. P.; Bratcher, M. S. J. Am.
Chem. Soc. 2005, 127, 6916–6917.
(41) Jyothish, K.; Zhang, W. Angew. Chem., Int. Ed. 2011, 50, 3435–3438.
(42) Gil-Ramirez, G.; Karlen, S. D.; Shundo, A.; Porfyrakis, K.; Ito,
Y.; Briggs, G. A. D.; Morton, J. J. L.; Anderson, H. L. Org. Lett. 2010,
12, 3544–3547.
(43) Meng, W. J.; Breiner, B.; Rissanen, K.; Thoburn, J. D.; Clegg,
J. K.; Nitschke, J. R. Angew. Chem., Int. Ed. 2011, 50, 3479–3483.
(44) Tashiro, K.; Yanagisawa, M.; Yamasaki, M.; Aida, T. J. Am.
Chem. Soc. 2007, 129, 11912–11913.
(45) Case, D. A. et al. AMBER 11; University of California,
San Francisco, 2010.
(46) Wang, J. M.; Wolf, R. M.; Caldwell, J. W.; Kollman, P. A.; Case,
D. A. J. Comput. Chem. 2004, 25, 1157–1174.
(47) Jakalian, A.; Bush, B. L.; Jack, D. B.; Bayly, C. I. J. Comput. Chem.
2000, 21, 132–146.
(48) Nishio, M.; Umezawa, Y.; Tsuboyama, S.; Takahashi, H.;
Uzawa, J. Tetrahedron 1999, 55, 10047–10056.
21001
dx.doi.org/10.1021/ja210418t |J. Am. Chem. Soc. 2011, 133, 20995–21001